Trying It on for Size

Online split testing gave Cabela’s the feedback needed to implement useful changes and avoid unproductive ones.

For retailers, one of the most convenient aspects of the online world is its flexibility. But they’ve also found that there’s an art to testing what works and what needs to change.

Vince Stephens, manager of Internet planning and analytics at Sidney, Neb.-based Cabela’s, an outdoor- gear retailer that accrued $1.8 billion in annual sales in 2005, said he learned the importance of testing after sitting in on a presentation by San Diego-based pet-store giant Petco, at the 2005 eTail conference in Palm Desert, Calif.

Stephens was surprised by how retailers were reaping major benefits from basic online tweaks and changes. Petco is a prime example. By making Web-page title, color, language and image adjustments, the pet center created a strategy that was positively influencing click-through rates.

Intrigued by Petco’s report, Cabela’s also wanted to optimize its complex, high-volume Web site with a solution that could quickly test different variations of content and functionality. But the retailer didn’t want to deplete time and dollars from its IT department.

The retailer turned to Boston-based SiteSpect. The company provides an automated, non-intrusive Web-based solution that allows retailers to create variations of site content, run split and multivariate tests on actual site visitors and analyze results in real time.

The solution, also called SiteSpect, enables retailers to test content changes to determine which will generate shopper purchases, join mailing lists or adopt self-service functions. Retailers can test content variations and track conversion behavior, without involving a company’s IT department.

For Cabela’s, Web analytics was a key part of the e-commerce strategy.

“Since our e-commerce team was pretty data-driven, we always recognized the value and importance of this kind of insight,” Stephens said. “But if someone came up with a good idea, we were challenged with how to act on it.”

Instead, Web upgrades were made based on brainstormed ideas and what appeared to be working in the industry.

“But split testing, or using two sets of Web-based services to different audiences, brought the ability to test some of our ideas and justify them with hard numbers,” he said.

Cabela’s began using the application to test the effectiveness of payment options, variations of shopping-cart solutions and merchandising features. As a result, the retailer started to see big improvements.

By testing different versions of written text on the page for example, Cabela’s improved its ability to up-sell products based on existing items in a visitor’s shopping cart, Stephens said. While testing this cross-selling strategy, the click-through rate increased 7%, and the average order value increased nearly 2%.

“The solution also pinpointed options that didn’t help the conversion rate, but would have taken a lot of time and money to implement,” he said. “We’ve also experienced a solid return on investment and significant savings by avoiding development expenses on changes that would not have yielded incremental sales.”

The solution alerted Stephens how a feature could have an effect on conversion rates, regardless of whether or not people actually used it.

Through a split test, Cabela’s compared product detail pages that showed instructional videos to pages without the feature. The results revealed a double-digit increase in item view-to-order conversion when the product video was shown.

“The information blew our minds,” he said. “The appearance of more product information seemed to affect customer confidence and purchasing behavior, even if they weren’t watching the video.”

Stephens said Cabela’s has used SiteSpect to perform some relatively sophisticated tests and will continue to use it as a proof-of-concept tool. But he also said he wants to refocus on some of the basics such as titles, images and colors in the near future in order to zero in on how they impact shopper behavior.

“We still think there are a lot of big wins in the simple changes,” he said. “So that’s where we’re headed next, back to the basics.”

Login or Register to post a comment.